基于EEMD和频带能量比特性的人与动物雷达微动信号辨识  

Micro-Vibration Distinguishment of Radar Between Humans and Animals Based on EEMD and Energy Ratio Characteristics

作　　者：殷悦 于霄[1] 吕昊 祁富贵 张自启[1] 王健琪[1] YIN Yue;YU Xiao;LV Hao;QI Fugui;ZHANG Ziqi;WANG Jianqi(School of Biomedical Engineering,The Fourth Military Medical University,Xi＇an Shannxi 710032,China)

机构地区：[1]第四军医大学生物医学工程学院,陕西西安710032

出　　处：《中国医疗设备》2018年第10期27-31,共5页China Medical Devices

基　　金：国家自然科学基金(81601567;61327805)

摘　　要：生物雷达是一种可穿透非金属介质、非接触探测到生命体的特殊雷达。它通过检测由生命体呼吸和体动等所引起的微动,从而探测生命体。对于生命体辨识的研究,目前多为运动状态下的目标动作辨识,而对静止状态的目标类型辨识仍然存在困难。为解决上述问题,本文根据生命体的回波信号特性,选择聚合经验模态分解处理非平稳的生命体微动信号。然后利用噪声和一般信号的自相关函数特性不同,重构目标微动信号,根据微动信号在参考频带上的能量占比大小辨识人与动物。数据处理结果表明,人和兔的频带能量比数值明显不同,差异具有统计学意义(P<0.001)。本方法在一定程度上可以区分人与兔目标,使静止状态下人与动物的辨识成为可能。Bio-radar is a kind of special radar which can penetrate nonmetallic medium and detect living bodies in non-contact situation. The radar can recognize living bodies through detecting micro-vibrations caused by respiration, body movement and so on. With regard to the recognition of the living body, most studies are about activity recognition in motion state. However, it still remains diffcult to identify motionless targets. To solve the above problem, we chose ensemble empirical mode decomposition method to analyze the non-stationary micro-vibration signal based on echo properties of living body. Then, in order to reconstruct micro-vibration signal, we utilized different characteristics of self-correlation function between the noise and general signal. Finally, we could distinguish humans from animals according to the energy ratio of micro-vibration signal in the reference band. Experimental results showed that the energy ratios of frequency band between humans and rabbits were obviously different and the difference was statistically significant （P〈0.001）. The method we proposed can distinguish between humans and rabbits to a certain extent, making it possible to recognize motionless humans and animals.

关 键 词：生物雷达 聚合经验模态分解 自相关函数 频带能量比 

分 类 号：TN957[电子电信—信号与信息处理]